Intra Prediction Filtering for Video Block Boundary Discontinuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video signal encoding and decoding, the discontinuity of block boundaries occurs when the prediction mode of a current block differs from that of a neighboring block, leading to increased difference values and reduced intra-prediction performance.
Innovation Solution
A method is introduced to filter the prediction block by selecting the prediction mode of the neighboring block, deriving a reference sample, and applying weights to generate an enhanced prediction block, thereby reducing boundary discontinuity and improving prediction performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra prediction is performed with different prediction modes for current block and neighboring block, then prediction accuracy for each block is improved, but block boundary discontinuity increases and difference values increase
Solution Approach 1:
The patent introduces a boundary filtering process as an intermediary step between prediction block generation and reconstruction. This filter uses reference samples from neighboring blocks to generate filtered prediction samples at block boundaries, mediating between different prediction modes and reducing discontinuity while preserving the benefits of mode diversity
Solution Approach 2:
The patent applies different processing to different regions of the prediction block. Boundary regions undergo filtering using reference samples from neighboring blocks, while interior regions use standard intra prediction. This local differentiation resolves the contradiction by addressing boundary discontinuity specifically without affecting overall prediction accuracy
2Stability of the object's composition
If boundary filtering is applied to reduce difference values, then block boundary continuity is improved, but encoding complexity increases
Solution Approach 1:
The patent performs boundary filtering using already-decoded reference samples from neighboring blocks before the current block is fully processed. This preliminary action reuses available data without requiring additional complex operations, reducing encoding complexity while maintaining boundary continuity
Solution Approach 2:
The boundary filtering process uses reference samples that are naturally available from the decoding of neighboring blocks. The system serves itself by reusing existing decoded data for boundary filtering, avoiding the need for separate complex boundary processing mechanisms
Data Source
AI summary
The present invention provides a method for decoding a video signal, which includes: generating a first prediction block according to a prediction mode of a current block; checking the prediction mode of a neighboring block adjacent to the current block; generating a second prediction block based on the prediction mode of the neighboring block; generating a final prediction block by applying a weight to the first prediction block and the second prediction block; and reconstructing the video signal by using the final prediction block.